Resveratrol enhances HBV replication through activating Sirt1-PGC-1α-PPARα pathway

The population of hepatitis B combined with a number of metabolic disorders is increasing significantly. Resveratrol (RSV) has been used as a preclinical drug for the treatment of the metabolic disorders. However, the impact of RSV on HBV replication remains unknown. In this study, the HBV-expressing hepatocelluar carcinoma cell line and mouse model created by hydrodynamic injection of viral DNA were used. We found that RSV activates Sirt1, which in turn deacetylates PGC-1α and subsequently increases the transcriptional activity of PPARα, leading to the enhanced HBV transcription and replication in vitro and in vivo. In addition, we found that this pathway is also required for fasting-induced HBV transcription. Taken together, this study identifies that RSV enhances HBV transcription and replication especially acting on the core promoter, which depends on Sirt1-PGC-1α-PPARα pathway. We conclude that RSV may exacerbate the progression of hepatitis B and that patients with hepatitis B infection should be cautious taking RSV as a dietary supplement.


RSV increases HBV replication in vitro and in vivo.
We first tested the toxicity of RSV in HepG2. 2.15 cells and determined the optimal dose for the later treatment. The 50% inhibitory concentration (IC50) of cell viability for RSV treatment was 106.3 μM and then the 50 μM was an effective concentration chosen for the following experiments of RSV (Fig. 1a). To explore the impact of RSV on HBV replication in HepG2.2.15 cells, equivalent numbers of cells were exposed to 50 μM RSV or control vehicle for 72 hours and HBV DNA in the culture medium was measured by real-time qPCR. We found that RSV treatment dose-dependently upregulated HBV DNA levels by nearly 8 times higher than the control (Fig. 1b). Same results were obtained in HepG2 (data not show) and Huh7 cells transiently transfected with 1.3mer HBV genomic DNA (Fig. 1c). We then established an HBV replication mouse model with hydrodynamic injection of 1.3mer HBV genomic DNA (10 μg/per mouse) according to a previous study 24 . After injection, the serum HBV DNA number in mice remained at a low level of [10 3 ] IU/ml for more than one month (Fig. 1d), which is consistent with the previous data 25 . After treatment of RSV (100 mg/kg/d) via daily oral gavage of RSV for 2 weeks, the serum HBV DNA levels were significantly increased and maintained at a higher level towards the end point of the drug administration (Fig. 1e), suggesting that RSV is a strong stimulator for HBV DNA replication.
RSV activates the HBV core promoter. To further investigate the mechanism by which RSV upregulates HBV replication, we measured the pregenomic RNA (pgRNA) and preC RNA in HepG2.2.15 cells by real-time qPCR, HBcAg expression by Western blot, and the HBsAg and HBeAg contents released into the culture medium by ELISA. We found that RSV treatment increased not only pgRNA and preC RNA levels by almost 6-fold and 4-fold respectively, but also HBeAg and HBcAg levels by nearly 5-fold and 3-fold respectively, with a slight increase in HBsAg expression (Fig. 2a), suggesting that the transcription of HBV DNA was activated by the RSV treatment, especially at the region of viral core promoter. Our in vivo experiments further supported these observations. RSV treatment (100 mg/kg/d) for 2 weeks increased HBsAg and HBeAg levels in circulation  (Fig. 2b,c), indicating that the viral core promoter plays a key role in the process. We then measured the activities of the preS1, preS2 and core promoters respectively in HepG2 cells. We found that while RSV treatment increased the core promoter activity by nearly 16-fold, it activated preS1 and preS2 promoters with much less potency (Fig. 2d). To further confirm the results in vivo, 6 μg of 1.3× HBV-Cp-Luc plasmid DNA was hydromatically injected into the mice. 72 hours later, the mice were treated with RSV for 20 minutes. The activity of the core promoter in mice liver was measured by the animal imaging system in real time. We found that RSV increased the luciferase signals in the liver by nearly 1.3-fold, compared with the vehicle control (Fig. 2e). These results suggest that RSV promotes HBV replication by activating the HBV core promoter. In vivo luciferase analysis of mice was observed after administration of water (control group, NC) or RSV (experimental group, Exp). Representative images for visualization of luciferase activity (left) and their quantitative analysis data (right) were shown. The gels have been run under the same experimental conditions. *p < 0.05 with n = 6 /group.  26 . To identify the transcription factor(s) involved in the activation of the core promoter by RSV, we knocked down various HBV-related transcriptional factors including Jun Proto-Oncogene (c-Jun), peroxisome proliferator-activated receptor, alpha/retinoid x receptor, alpha (PPARα/RXRα), CCAAT/Enhancer binding protein (C/EBP), alpha (C/EBPα), farnesoid x -activated receptor (FXRα), and hepatocyte nuclear factor 4, alpha (HNF4α) in HepG2.215 cells by small interfering RNAs (siRNAs). Real-time qPCR data demonstrated that knockdown of PPARα or RXRα abolished RSV-induced HBV replication ( Fig. 3a), while knockdown of the others didn't affect supernatant HBV DNA levels (Fig. 3b). To examine whether PPARα /RXRα could regulate the HBV core promoter, we knocked down PPARα or RXRα by siRNA in HepG2 cells following pGL3-Cp transfection and RSV treatment. Evidently, knockdown of either PPAR or RXR reduced the luciferase signal of the core promoter (Fig. 3c). To further evaluate the activity of PPARα / RXRα affected by RSV treatment, we measured the expression of PPARα /RXRα target genes by RT-qPCR in HepG2.2.15 cells and the mouse liver. RSV stimulated the expression of PPARα target genes promoting fatty acid oxidation (FAO), such as key mitochondrial medium chain acyl-CoA dehydrogenase (MCAD), microsome cytochrome P450 in FAO and mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase (mHMG-CoAs) in ketogenesis (Fig. 3d). Taken together, the activation of PPARα /RXRα on core promoter plays a key role in RSV-induced HBV transcription.
Sirt1/PGC-1α is required for RSV-induced HBV replication. RSV can act as a positive regulator of sirtuin 1 (Sirt1) to improve the metabolic process of aging, hepatic steatosis and diabetes [27][28][29][30][31] . To examine whether Sirt1 is also involved in RSV-induced HBV replication, we measured the expression of endogenous Sirt1 as well as its activity in HepG2.2.15 cells and mouse liver. We found that RSV not only increased Sirt1 mRNA and protein levels ( Fig. 4a,b), but also enhanced the ratio of NAD/NADH (Fig. 4c), consistent with the previous reports 32, 33 .
To examine whether Sirt1 contributes to RSV-induced HBV replication, we measured HBV DNA in the culture media from Sirt1 knockdown cells (Fig. 4d). Knockdown of Sirt1 significantly decreased the HBV DNA levels by almost 80% compared to the control (Fig. 4e). The activation of HBV core promoter by RSV was also substantially blocked by Sirt1 knockdown (Fig. 4f). In addition, nicotinamide (NAM), a Sirt1 activity inhibitor, also dramatically reduced RSV-stimulated HBV core promoter activity (Fig. 4g). These data demonstrated an essential role of Sirt1 in RSV-stimulated HBV replication.
Previous studies have demonstrated that in hepatic lipid metabolism, Sirt1 can activate PPARα via deacetylating peroxisome proliferator-activated receptor gamma coactivator 1, alpha (PGC-1α ) 34 . To analyze whether PGC-1α is a signal that links Sirt1 to PPARα /RXRα in RSV-induced HBV replication, we examined the PGC-1α acetylation in HepG2.2.15 cells. RSV treatment significantly decreased the acetylation level of PGC-1α protein in HepG2.2.15 cells (Fig. 5a) and increased the interaction between endogenous Sirt1 and PGC-1α (Fig. 5b). Pretreatment of nicotinamide (NAM), a Sirt1 activity inhibitor, restored the acetylated PGC-1α level to that of the untreated group (Fig. 5a). To determine whether PGC-1α is required for RSV-induced HBV replication, we knocked down PGC-1α by siRNA in HepG2.2.15 cells. PGC-1α deficiency caused a significant decrease of HBV DNA in the culture medium compared to the control (Fig. 5c). Similarly, knockdown of PGC-1α significantly reduced RSV-stimulated core promoter activity in HepG2 cells (Fig. 5d). In sum, these findings indicate that Sirt1 activated by RSV directly regulates PGC-1α acetylation, resulting in RSV-induced HBV replication.
Activation of PPARα is required for fasting-induced HBV replication. Calorie restriction has a similar effect on energy metabolism to that of RSV treatment 35,36 . It was reported that fasting can induce HBV gene expression through PGC-1α 37 . Given that the fact that PPARα 's transcriptional activity can be activated by fasting 38,39 and the essential role of PGC-1α /PPARα axis in mediating RSV-induced HBV replication, we reckoned that PPARα may be required for fasting-triggered HBV replication. To test this idea, the mice were starved for different time periods after injection of 1.3 × HBV-Cp-Luc plasmid DNA. The core promoter activity, assessed by the luciferase intensity, was time-dependently increased in all fasted mice (Fig. 6a,e). To verify the role of (g) Sirt1 antagonist NAM markedly blocked RSV-stimulated activity of the viral core promoter. HepG2 cells were cotransfected with pGL3-Cp and pRL-TK plasmids for 24 hours and cells were then treated with RSV (50 μM) or RSV (50 μM) plus NAM (10 mM) for 72 h. The gels have been run under the same experimental conditions. All experiments were repeated at least three times with consistent results. Bar graphs represent the means ± SEM, n = 3 (*p < 0.05).
PPARα in the process, the hepatic expression of PPARα targeted genes including MCAD, cytochrome P450 and mHMG-CoAs was measured by RT-qPCR. Starvation up-regulated the mRNA levels of all the genes measured ( Fig. 6b), similar to those caused by RSV treatment (Fig. 3d). In addition, GW6471, a specific PPARα inhibitor, significantly blocked the enhanced activity of the HBV core promoter by fasting in the time course of starvation (Fig. 6d,e), suggesting an essential role of PPARα activation in fasting-induced HBV transcription.

Discussion
RSV has attracted tremendous attentions in the past years due to its beneficial effects in a number of diseases (e.g. metabolic or cardiovascular diseases, aging, cancer ) exerted by its protective functions such as antioxidant properties 15,40 , anti-inflammatory 41 and anti-proliferative functions 42 respectively. Although RSV was found to have effects on several viral infections, this study is the first time to demonstrate that RSV actually induces HBV replication in vitro and in vivo. Given the fact that there are increasing successful cases of preclinical trials of RSV treatment on a variety of human diseases 43,44 , our data raises a serious concern about the risk of possible side effects of RSV, such as treatment of RSV in patients with metabolic disorders simultaneously infected with HBV. In addition, RSV exists in a number of fruits such as red grapes (50-100 mg/g) 45 and drinks such as red wine (0.2 mg/l to 5.8 mg/l) 46 at a notable concentrations. Patients with HBV infection, especially HBV carriers and occult infection patients, would have a risk of HBV recurrence when drinking over-dosed red wine. Given the drinking culture in large populations worldwide, additional studies would be required to determine the guideline amount for red wine drinking in HBV patients.
Our data are consistent with previous finding that RSV is considered as an activator of Sirt1 47 , a member of the mammalian sirtuins or HDAC class III. Ren 49 . The discrepancy between the two studies discussed above may be partially due to different cell models and experimental conditions used. In our study, unlike Act3, RSV alone displays a strong ability to promote HBV replication in a Sirt1-PGC-1α and PPARα /RXRα -dependent manner. Sirt1, PGC-1α and PPARα / RXRα proteins appear to constitute a sub-network that modulates the HBV core promoter activity and viral transcription. RSV-induced Sirt1 activation mediates HBV transcription, and inhibition of Sirt1 by NAM attenuates this effect, suggesting that activation of Sirt1 is required for RSV-induced HBV replication. Given the limited effects of Sirt1 overexpression or activation by Act3 on the HBV transcription, the most likely explanation is that RSV has both Sirt1-dependent and independent functions 50 , which can simultaneously activate other signals to augment the function of this sub-network, such as activation of PPARα /RXRα through other pathways.
Fasting induces metabolic responses that allow mammals to survive for a long period of energy deprivation. Dramatic changes in gluconeogenesis and fatty acid oxidation are prominent features of the energy-metabolic responses. Shlomai A et al. demonstrated that PGC-1α controls HBV replication through nutritional signals and interestingly, PGC-1α co-activates HNF4α , a key enzyme of gluconeogenesis, to promote HBV replication 37 . However, we think that the role of HNF4α is not conclusive due to lack of loss of function study to demonstrate the necessity of HNF4α in this process. In our study, although the expression of gluconeugenic genes was upregulated after 7-hour fasting, there was no similar change in the gene expression in the time points of prolonged fasting (Fig. 6c), suggesting that the activation of gluconeogenesis is not associated with fasting-induced HBV viral replication. Instead, we found that genes involved in FAO were closely associated with fasting-induced HBV transcription and replication. This was further confirmed by our experiments where inhibition of PPARα , a key enzyme of FAO, alleviated fasting-induced HBV transcription in vivo, providing strong evidence that FAO plays an important role in HBV replication in the hosts exposed to nutrient/energy deficiency.
In summary, this study demonstrates that RSV has a strong ability to enhance HBV replication through its core promoter. As shown in Fig. 7, RSV activates Sirt1 and enhances HBV replication in a PGC1α -PPARα /RXRα dependent mechanism, which resembles the pathway of fasting-induced FAO and HBV replication. Supplement of HBV patients with RSV presents a potential risk of hepatitis B recurrence.

Animal experiments.
Our animal studies were conducted in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals according to the regulation in the People's Republic of China. The protocol was approved by the Committee on the Ethics of Animal Experiments of Zhejiang University (Approval No. X1201231). All procedures were made to minimize suffering. C57BL/6 mice (male, 6-8 weeks old), obtained from Shanghai Laboratory Animal Center (Shanghai, China), were used for animal experiments. For the whole HBV analysis, ten micrograms of HBV plasmid DNA (pHBV) in a volume of PBS equivalent to 10% of the mouse body weight were injected via tail vein on day 0 according to the method 24 . Twenty eight days after the hydrodynamic injection, mice were divided into two groups as follows: RSV group and H2O group. Mice received either RSV (100 mg/kg mouse body weight) or vehicle (water) by oral gavage every day for 2 weeks. The chosen dose was based on a previous study 51 . RSV was re-suspended in water. Mice were sacrificed at 35 and 42 days after injection. Liver tissues and sera were collected for pathologic and biochemical analysis.
For animal imaging experiments, mice were injected with six micrograms of the 1.3× HBV-Cp-Luc construct and analysis was performed 72 h after the hydrodynamic injection 37 . After mice received an intragastric administration of RSV (100 mg/kg/d) or vehicle (water) for 20 minutes, mice were anesthetized with isoflurane, and D-luciferin potassium salt (122796, Perkin Elmer, USA) at 150 mg/kg was then intraperitoneal injection (i.p.) injected. Visualization of luciferase activity was performed on the Lumazone FM1024 equipment (Nippon Roper, Japan). Data analysis including quantification was performed by using the SlideBook ™ 4.0 software (Intelligent Imaging Innovations, USA). Luciferase activity of every animal was quantified and expressed as the ratio relative to the baseline activity.
For starvation experiments, male littermates were separated into individual cages at the beginning of each fasting experiment. Fasting was initiated at 5:00 p.m. Mice injected with the 1.3x HBV-Cp-Luc plasmid at 72 h prior to fasting were divided into two groups. The experimental group (n = 6) was fasted (free access to water was allowed) at indicated times (0, 7, 24, 48 h) and allowed to a subsequent 24-hour refeeding. The control group (n = 6) was let for continuous free feeding. The luciferase activity of core promoter was visualized at indicated times. Meanwhile, another 30 mice (divided into 5 groups: 0 h, fasting 7 h, fasting 24 h, fasting 48 h, refed, n = 6) treated as indicated above were killed and liver tissues were collected and immediately frozen in liquid nitrogen for further analysis. To investigate the effect of inhibiting activity of PPARα on the viral core promoter under nutritional deprivation, mice with HBV-Cp-Luc plasmid were administered acute intraperitoneal injection of GW6471 (1 mg/kg, a PPARα antagonist) 30 min prior to defined fasting time points. The dose was chosen based on the previous pilot study 52 . Control animals received the same volume of the vehicle (1% aqueous solution of DMSO). Luciferase activity of every animal was quantified at all time points.
For the detection of pgRNA and preC RNA, the primers used were based on the previous pilot study 54 . In brief, the cDNA product was used in each of three separate amplification reactions with BC1 (5′ -GGAAAGAAGT CAGAAGGCAA) as the common 3′ antisense primer and 5′ sense primers: (1) PCP (5′ -GGTCTGCGCACCAGCACC) for the specific detection of preC RNA transcripts, (2) PGP (5′ -CACCTCTGCCTAATCATC) for monitoring total CP-directed RNA transcription (pgRNA plus preC RNA), and (3) M3 (5′ -CTGGGAGGAGTTGGGGG AGGAGATT) for detecting residual HBV DNA contamination. The levels of pgRNA transcripts were calculated by subtracting preC RNA levels from total CP-directed transcription.

Protein-protein interaction analysis and Western blot. Endogenous protein-protein interaction in
cells was examined by co-immunoprecipitation experiments using anti-Sirt1 and anti-PGC-1α antibodies. Cells were lysed with Cell Lysis Buffer (1× ) (#9803, Cell Signaling Technology, USA) containing 1 mM PMSF, 10 mM nicotinamide and 10 μM TSA. Lysates were centrifuged (13,000g, 4 °C, 10 min) and the supernatants were used for immunoprecipitation. 50 μl of fresh protein G magnetic beads (#LSKMAGG02, Millipore, USA) were added and incubated with 1 μg (2 μL) of anti-Sirt1 or 2.4 μg (12 μl) of anti-PGC-1α for 10 minutes with continuous mixing at room temperature. Cell lysate samples (400 μg) and the immobilized capture antibody were then incubated at 4 °C with continuous mixing overnight. Immunocomplexes were washed several times, denatured with 80 μl 2× Laemmli sample buffer (10 min, 95 °C) and then analyzed by Western blot. The immunoprecipitates were separated by SDS-PAGE and immunoblotted using antibodies against Sirt1 and PGC-1α .

PGC-1α acetylation assays. PGC-1α acetylation level was measured by immunoprecipitation of PGC-1α
followed by Western blot using anti-acetyl-lysine antibody. PGC-1α and acetylation levels were assayed using specific antibodies for PGC-1α and acetyl-lysine.
Luciferase ashsay. HepG2 cells in 96-well plates containing 2.0 × 10 4 cells were transiently transfected with the reporter vector (pGL3-Cp, pGL3-S1p and pGL3-S2p) by using Lipofectamine 3000 according to the manufacturer's instructions. Transfection mixtures for each well comprised 100 ng of promoter reporter plasmid and 10 ng of plasmid pRL-TK, serving as an internal control to normalize the transfection efficiency. Six hours after transfection, RSV was added to the medium as indicated and cells were incubated for 3 days. Then Firefly and Renilla luciferase activities were measured by using a Dual-Glo ® Luciferase Assay System kit (E2940, Promega, USA) according to the manufacturer's instructions. The luciferase activity was determined on a GloMax microplate luminometer (Promega, USA). NAD+/NADH ratio assay. HepG2.2.1.5 cells were grown to 55% confluency in a 6 cm 2 tissue culture plate.
Cells were subsequently treated with media containing DMSO or RSV (50 μM). After 24 hours of treatment, cells were lifted with trypsin, washed twice with cold PBS and pelleted through centrifugation. NAD + /NADH ratio was performed using the EnzyChrom NAD/NADH Assay Kit (E2ND-48, Bioassay Systems, USA). NAD and NADH contents were normalized by protein concentrations in cell lysates. Detection of serum HBV antigen and DNA. Serum HBsAg and HBeAg levels were detected automatically by Abbott i2000SR using the Architect HBsAg and HBeAg Reagent kits (Abbott Diagnostics, Abbott Park, IL, USA). Serum HBV DNA copies were measured using the Fluorescence Quantitative PCR Detection Kit for Hepatitis B Virus DNA (ACON Biotech Co. Ltd, Hangzhou, China). The HBsAg, HBeAg and HBV DNA in the culture medium were measured similarly. All of these assays were conducted following the manufacturers' instructions.

Immunohistochemistry.
Statistics. Data were analyzed using GraphPad Prism v5.0a (GraphPad Software, Inc., SanDiego, USA). Data were presented as mean ± SEM. Statistical significance of the differences was determined using Student t test. Differences were considered significant when P < 0.05.